Content-based Movie Recommender System Using Keywords and Plot Overview

Big tech companies like Amazon, Netflix and Google have tons of data and are still successful in providing specific products and services correctly as per user requirements. This is made possible by the recommendation algorithms that feed on the data we provide, in turn, enabling them to produce accurate results. Movie recommendation systems aspire to help cinema geeks by proposing movies of their penchant, devoid of them needing to do the standard long and arduous method of selecting from huge sets of movies that go up to millions and is onerous and frustrating. In this paper, we aspire to diminish human endeavor by recommending them movies based on their interests. To resolve such troubles, we have built a model using a content-based approach. The idea behind this model is to recommend a movie based on descriptions of movies. Using the movie “GoldenEye” as an example, we obtained the result as “Skyfall” with a similarity score of 66.73% using CountVectorizer, 13.14% using Jaccard Recommender, 14.34% using TF-IDF Keywords, 9.87% using TF-IDF Plot Overview and 71.9% using Google Form responses.